HAT ‐ ECG : Hybrid autoencoder‐transformer architecture for ECG arrhythmia classification
Physiological Reports · 2026-06 · 14 (12)
DOI: 10.14814/phy2.70990 ↗At a glance
Journal article in Physiological Reports, published online on 2026-06-23. Written by Shahin Sharbaf Movassaghpour and 2 co-authors. Institution: Department of Computer Engineering, Ta.C., Islamic Azad University, Tabriz, Iran. Major MEDLINE topics: Arrhythmias, Cardiac, Electrocardiography, Signal Processing, Computer-Assisted. Cited 1 times so far (Crossref); 52 references. The full text is freely available (open access, CC BY 4.0).
Abstract (original language)
Accurate and efficient analysis of electrocardiogram (ECG) signals is essential for early detection of cardiac arrhythmias. However, many deep learning models suffer from limited generalization to unseen patients, reduced interpretability, and high computational demands. To address these challenges, we propose HAT‐ECG, a hybrid Autoencoder‐Transformer architecture that integrates unsupervised feature learning with attention‐based temporal modeling. The convolutional autoencoder extracts compact, noise‐robust latent representations of ECG beats, while the Transformer's Multi‐Head Attention mechanism adaptively focuses on diagnostically relevant waveform segments. The model was evaluated on three public datasets: MIT‐BIH, INCART, and the independent PTB Diagnostic ECG Dataset (290 subjects, entirely outside the original datasets). Under standard beat‐wise splitting for benchmarking, HAT‐ECG achieved state‐of‐the‐art accuracies of 99.91% (MIT‐BIH 5‐class), 99.69% (MIT‐BIH AAMI), 99.15% (INCART 3‐class AAMI), and 98.45% (PTB 2‐class). Critically, under strict patient‐wise splitting on MIT‐BIH (completely disjoint training and test patients), the model maintained a realistic accuracy of 90.81% (F1‐score 92.61%), with strong generalization to new patients. With only 0.021 GFLOPs, HAT‐ECG offers an excellent balance of high performance, interpretability, and efficiency, making it highly suitable for real‐time wearable and edge‐device cardiac monitoring. This work advances deep learning for intelligent and deployable arrhythmia classification by combining accuracy, cross‐patient generalization, and computational minimalism.
Subjects
MeSH terms (MEDLINE)
Author keywords
Authors and institutions
- Department of Computer Engineering, Ta.C., Islamic Azad University, Tabriz, Iran.
ORCID profiles: Shahin Sharbaf Movassaghpour, Masoud Kargar, Ali Bayani
Cite
APA
Sharbaf Movassaghpour, S., Kargar, M., & Bayani, A. (2026). HAT ‐ ECG : Hybrid autoencoder‐transformer architecture for ECG arrhythmia classification. Physiological Reports, 14(12). https://doi.org/10.14814/phy2.70990
Vancouver
Sharbaf Movassaghpour S, Kargar M, Bayani A. HAT ‐ ECG : Hybrid autoencoder‐transformer architecture for ECG arrhythmia classification. Physiological Reports. 2026;14(12). doi:10.14814/phy2.70990
BibTeX
@article{SharbafMovassaghpour2026,
author = {Sharbaf Movassaghpour, Shahin and Kargar, Masoud and Bayani, Ali},
title = {HAT ‐ ECG : Hybrid autoencoder‐transformer architecture for ECG arrhythmia classification},
journal = {Physiological Reports},
year = {2026},
volume = {14},
number = {12},
doi = {10.14814/phy2.70990}
}
EUVIMED